Project 3 will be directed towards the origins of BE and response to ablative therapy, and represents an extension of the hypotheses from Projects 1 and 2, specifically that chronic inflammation and activation of stem/progenitor cells are key factors in the development of BE. Data from histopathologic studies in humans and from our L2-IL-1 beta mouse model suggest that CLE (independent of goblet cells) may represent tissue at risk for neoplastic progression. However, there are currently no validated biomarkers of CLE as tissue with neoplastic potential. In patients who progress, endoscopic therapy with radiofrequency ablation (RFA) can eradicate both dysplasia and intestinal metaplasia. We have found that 25-35% of ablated patients develop IM and even dysplasia at the GE junction. We therefore propose to study this patient population as a novel human model for the development of BE through evaluation of biomarkers for recurrent IM. Using this same cohort, we will evaluate pre-treatment markers that predict short- and long-term response to RFA, as the ability to identify non-responders prior to treatment would allow better targeting of resources to those who would predictably respond to the therapy. In order to address these issues, we propose three aims: 1) To determine the site of origin of BE based on gene expression profiles in humans. We will extend the findings from the L2-IL-1beta mouse model to humans by comparing gene expression profiles of Barrett's and cardia tissue in BE patients and controls. We will evaluate specific markers of stem cell activation as well as new biologically-based markers identified by genomic analyses, as potential new biomarkers for CLE and BE; 2) To determine whether stem cell markers predict recurrence of BE after ablation therapy. We hypothesize that the recurrence of IM after ablative therapy is secondary to stem cell differentiation into an intestinal phenotype in the cardia, and that central obesity may impact this process; 3) To determine biomarkers which can predict response to ablation therapy in patients with BE. We will prospectively follow a large cohort of BE patients with HGD/EAC who undergo RFA. We will determine if pi6 status impacts efficacy will explore the possibility that epigenetic changes represent potentially modifiable predictors of non-response to therapy.